Llama 3.2 11B

Training compute

5.8×10²³ FLOP

Parameters

10.6B

Published

Sep 24, 2024

Llama 3.2 11B is an AI model developed by Meta AI (United States), first published in September 2024. It works in the multimodal, vision and language domain, on tasks such as visual question answering, image captioning and object detection.

Training it took an estimated 5.8×10²³ FLOP of compute (estimation method: hardware). The model has 10,600,000,000 parameters. Training ran on NVIDIA H100 SXM5 80GB.

Access: Open weights (restricted use). Its weights are openly available. It is built on top of Llama 3.1-8B. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Meta AI
Country of organization: United States
Domain: Multimodal, Vision, Language
Task: Visual question answering, Image captioning, Object detection
Training compute: 5.8×10²³ FLOP
Compute estimation method: Hardware
Parameters: 10,600,000,000
Training hardware: NVIDIA H100 SXM5 80GB
Chip-hours: 246.1K
Model accessibility: Open weights (restricted use)
Open weights: Yes
Base model: Llama 3.1-8B
Epoch confidence: Confident